Amazon Prime Day generated data that Amazon will mine for insights on new strategies for sales promotion. This high-velocity marketing campaign was an attempt to etch into consumer's minds an alternative to Black Friday. It was one of the first experimental marketing campaigns of this type—it won't be the last.
The media commentary on the event concluded that Amazon Prime Day fell short of expectations. Delivery delays and the perception that the product choices smacked of a clearance sale caused a consumer backlash, most notably on Twitter, inspiring the hashtag #PrimeDayFail. The revenue generated was below its potential because inventory depleted sooner than expected. Overall, Prime Day exposed weaknesses in Amazon's operations.
Jeff Gothelf, founding partner and lead evangelist of Neo, as well as coauthor with Josh Seiden of the upcoming book, Sense and Respond: Turning Your Organization Into an Innovation Machine, has a positive outlook. "Amazon is a continuously learning company. Just because something 'failed' the first time, they'll learn from it and do it better the second time around," he says.
I spoke to consultants, researchers, and industry experts to gain more insight into the causes of fragile relationships between operations, development, and marketing, and how DevOps methodologies can help manage these relationships for a more agile response to market needs. These experts advise companies on the roles that collaboration, automation, and metrics development play in ensuring that all three synchronize their activities for speedy responses.
Aligning marketing, development, and operations
Real-time marketing is fraught with the risk of a rocky journey as product development, operations, and marketing strain to synchronize their activities with only a hazy view of their fluid markets. The experience with Amazon Prime Day "shows an all-too-common misalignment between an organization's product teams and its marketing teams, which often have competing incentive structures," Gothelf explains.
The iterative nature of agile development practices should help teams execute on milestones in fast, incremental cycles that help their businesses cope with uncertainties in a fluid market. The problem is the absence of a comprehensive structure: marketing would like to add features as it gains visibility into consumer needs, but development wants to see a wireframe or vice versa. The belief is that this structure is essential to execution and that incremental operational changes could go awry if they're not consistent with an overarching architecture, especially in larger companies.
Comprehensive agile collaboration is needed
The solution is an agile management style that strengthens collaborative processes; blends marketing, development, and operations functions; and creates interdependence among teams for cohesive action. "This means that members of the marketing team are also members of the product development team. They attend stand-ups, do research with the product teams, and build evolutionary launch plans that refine as the product takes shape," says Gothelf.
Collaboration sets the stage for the integration and automation of the processes needed to speed up market responses. "APIs are the keystone for integrating analytics, internal business processes, and user experience for real-time marketing," says Michael Schrage, a research fellow at the MIT Center for Digital Business and advisor to several Fortune 500 companies. "DevOps mistakenly sees APIs as the realm of software development alone and neglects communication of the varied benefits, such as latencies, customization, and peak demand fulfillment they provide to marketing," Schrage explains. He coined the term "Waterfall APIs" to assert the importance of structure. "The organizing principle is use cases which in turn will call for service level agreements, set the parameters for wireframing and inventory management," Schrage adds. "Hierarchy is replaced by collaboration in this world, but architecture is still crucial."
Experiments expose fault lines
Dynamic pricing and promotions are increasingly recognized as a way to take advantage of opportunities that previously were left on the table due to slow responses. Now marketing wants to seize opportunities as soon as they become available and expect accelerated responses from their development and operations divisions. This will accentuate fault lines across development, operations, and marketing as impromptu offers and price discounts allow companies to seize sales opportunities. According to a study by RSR Research, 59 percent of retail companies believe that technologies for promotions planning, forecasting, and management are very valuable, while only 39 percent are satisfied with their existing deployments of these technologies. Similarly, regular price planning, forecasting, and management are seen as valuable by 61 percent of respondents, while only 41 percent of respondents are satisfied with their existing deployments.
Companies that want to make gains using market experiments are also more aware of the need for adaptive business processes. The number of companies that view business process analysis as important rose to 43 percent in 2015, from 26 percent the year before.
Connecting marketing with operations
Process changes begin with some of the basics of actionable information flows, from marketing to back-end operations. "Most retail companies are not able to correlate their aggregate sales performance of each marketing campaign with their segment, geographical, and channel level outcomes in real time," says Paula Rosenblum, retail analyst and managing partner at RSR Research.
Development, operations, and marketing need distinct metrics on campaign performance before they can pinpoint the causes of shortfalls. "The return on investment is not measured at a granular level for each marketing campaign and corresponding expected performance metrics for individual departments. As a result, these departments have no incentive to communicate with each other," Rosenblum says.
Tying metrics to business goals
In the final analysis, these metrics have to be tied to the business goals of the company, or they will lead to undesired outcomes. Amazon Prime Day achieved its sales goals but disappointed customers, many of whom will likely migrate to emerging competitors, such as Jet.com. "A strategic way to select a metric is to look at primarily the customer lifetime value and associated measures like customer retention and satisfaction, not just sales performance, which can only be effectively measured over several campaigns," says Gary Katz, chairman of Marketing Operations Partners.
"It's also important to look at leading indicators and not lagging ones, in order to be able to influence the outcomes with help from predictive models" Katz adds. "The nitty-gritty of measurements and model development is best left to the data science specialists who continue to improve predictive analytics."
In order to overcome the tensions between development, operations, and marketing, a root-and-branch transformation is needed before progress can be made. A preoccupation with the immediate marketing campaign won't improve performance. It's time to look at the issue holistically. This will require a review and restructuring of incentives, communication and information flows, metrics, and analytics.
Keep learning
Take a deep dive into the state of quality with TechBeacon's Guide. Plus: Download the free World Quality Report 2022-23.
Put performance engineering into practice with these top 10 performance engineering techniques that work.
Find to tools you need with TechBeacon's Buyer's Guide for Selecting Software Test Automation Tools.
Discover best practices for reducing software defects with TechBeacon's Guide.
- Take your testing career to the next level. TechBeacon's Careers Topic Center provides expert advice to prepare you for your next move.
